Peak Detection System and Method for Calculation of Signal-Derived Metrics

Abstract

The invention in at least one embodiment includes a method for providing a beat determination based on signals from a plurality of medical sensors in substantially real time, said method including: receiving in at least two analysis modules at least one signal (including an ECG signal and/or a pulse signal) from at least one medical sensor such that at least one analysis module receives one ECG signal and at least one analysis module receives one pulse signal; in each analysis module processing one signal through a plurality of detection algorithm modules to produce at least one quantitative decision, fusing the quantitative decisions together with a fusion module, and outputting at least one detected peak from the fusion module to an aggregation module; and producing from the aggregation module a series of detected peaks based on detected peaks received from the analysis modules. In a further embodiment, the detected peaks produced by the aggregation module are used to determine heart rate variability/complexity. At least one embodiment includes a system for performing any of the above methods.

What is claimed is: 
     
         1 . A method for providing a beat determination based on signals from a plurality of medical sensors attached to one patient in substantially real time, said method comprising:
 receiving in at least two analysis modules at least one signal from at least one medical sensor, wherein the signal includes at least one of an ECG signal and a pulse signal such that at least one analysis module receives the one ECG signal and at least one analysis module receives the one pulse signal;   in each analysis module
 processing the one signal through a plurality of detection algorithm modules with each detection algorithm module producing at least one quantitative decision, 
 fusing the quantitative decisions from the detection algorithm modules together with a fusion module, and 
 outputting at least one detected peak from the fusion module to an aggregation module in communication with the at least two analysis modules; and 
   producing from the aggregation module a series of detected peaks representing heartbeats of the patient based on a series of detected peaks received from the analysis modules.   
     
     
         2 . The method according to  claim 1 , wherein producing the series of detected peaks is further based on a beat signal quality index for each analysis module with the analysis module with the highest beat signal quality index being selected to supply the detected peaks for output by the aggregation module. 
     
     
         3 . The method according to  claim 2 , wherein the beat signal quality index equals the beats detected by that analysis module over a predetermined time period divided by an average of the total number of beats detected by all of the analysis modules over the predetermined time period. 
     
     
         4 . The method according to  claim 1 , wherein fusing the quantitative decisions together includes
 receiving quantitative decisions in a buffer of the fusion module from the decision algorithm modules in communication with the fusion module;   determining whether quantitative decisions have been received from all of the decision algorithm modules;   determining whether a most recently received quantitative decision is a second consecutive decision from one detection algorithm module;   when either determination is true,
 fusing the received quantitative decisions, and 
 determining an existence of a peak and outputting any peak that exists; and 
   when neither determination is true, waiting for the next quantitative decision to be received from the detection algorithm modules.   
     
     
         5 . The method according to  claim 4 , wherein fusing the quantitative decisions occurs after elapse of a refractory period after the last peak is determined to exist. 
     
     
         6 . The method according to  claim 1 , wherein each detection algorithm module
 pre-processes the one signal to at least one of filter the signal and extract features from the signal to provide at least one processed signal, and   analyzes the at least one processed signal to detect whether a beat is present in the processed signal.   
     
     
         7 . The method according to  claim 6 , wherein at least one analysis module includes at least two detection algorithm modules that share at least one processed signal. 
     
     
         8 . The method according to  claim 1 , wherein aggregating by the aggregation module includes
 receiving detected peaks in a buffer from the analysis modules;   determining whether detected peaks have been received from all of the analysis modules;   determining whether a most recent received detected peak is a second consecutive detected peak from one analysis module;   when either determination is true,
 fusing the received detected peaks, and 
 determining an existence of a peak and outputting any peak that exists; and 
   when neither determination is true, waiting for the next detected peak to be received from the analysis modules.   
     
     
         9 . The method according to  claim 8 , wherein fusing the received detected peaks by the fusion module includes analyzing the detected peaks using centroid analysis to detect output agreement. 
     
     
         10 . The method according to  claim 9 , wherein output agreement is based on K-nearest neighbor. 
     
     
         11 . The method according to  claim 1 , further comprising:
 receiving the series of detected peaks from the aggregation module into a heart rate complexity module;   determining heart rate complexity by the heart rate complexity module based on the received detected peaks; and   triggering an alarm when the heart rate complexity is outside a predetermined range.   
     
     
         12 . The method according to  claim 11 , wherein the heart rate complexity based on an entropy measure. 
     
     
         13 . The method according to  claim 1 , further comprising selecting the analysis module with a highest best signal quality index using a signal validation model. 
     
     
         14 . The method according to  claim 1 , wherein the receiving analysis module is selected based on the detection algorithm modules present in the receiving analysis module being configured to process the at least one signal. 
     
     
         15 . The method according to  claim 1 , wherein each detection algorithm module is configured to use different logic to detect a beat in the input signal as the quantitative decision. 
     
     
         16 . The method according to  claim 1 , wherein the plurality of detection algorithm modules in each analysis module numbers at least four. 
     
     
         17 . The method according to  claim 1 , wherein the plurality of medical sensors includes at least one ECG lead and a pulse oximeter. 
     
     
         18 . A method for providing a beat determination based on signals from a plurality of medical sensors attached to one patient in substantially real time, said method comprising:
 receiving in at least two analysis modules at least one signal from at least one medical sensor, wherein the signal includes at least one of an ECG signal and a pulse signal such that at least one analysis module receives the one ECG signal and at least one analysis module receives the one pulse signal;   in each analysis module
 processing the one signal through a plurality of detection algorithm modules with each detection algorithm module producing at least one quantitative decision, 
 fusing the quantitative decisions from the detection algorithm modules together with a fusion module, and 
 outputting at least one detected peak from the fusion module to an aggregation module in communication with the at least two analysis modules; and 
   producing from the aggregation module a series of detected peaks representing heartbeats of the patient based on a series of detected peaks received from the analysis module having a highest beat signal quality index for each analysis module with the analysis module with the highest best signal quality index being selected to supply the detected peaks for output by the aggregation module, and   wherein the beat signal quality index equals the beats detected by that analysis module over a predetermined time period divided by an average of the total number of beats detected by all of the analysis modules over the predetermined time period.   
     
     
         19 . The method according to  claim 18 , wherein
 fusing the quantitative decisions together includes
 receiving quantitative decisions in a buffer of the fusion module from the decision algorithm modules in communication with the fusion module; 
 determining whether quantitative decisions have been received from all of the decision algorithm modules; 
 determining whether a most recent received quantitative decision is a second consecutive decision from one detection algorithm module; 
 when either determination is true,
 fusing the received quantitative decisions, and 
 determining an existence of a peak and outputting any peak that exists; and 
 
 when neither determination is true, waiting for the next quantitative decision to be received from the detection algorithm modules, 
   fusing the quantitative decisions occurs after elapse of a refractory period after the last peak is determined to exist,   each detection algorithm module
 pre-processes the one signal to at least one of filter the signal and extract features from the signal to provide at least one processed signal, 
 analyzes the at least one processed signal to detect whether a beat is present in the processed signal, and 
   at least two detection algorithm modules in at least one of the at least two analysis modules share at least one processed signal.   
     
     
         20 . A computer program product for detecting peaks in a plurality of biomedical signals, the computer program product comprising:
 a computer readable storage medium having encoded thereon:   first program instructions executable by a processor to cause the processer to receive a plurality of biomedical signals originating from a plurality of medical sensors;   second program instructions executable by a processor to cause the processer to assign each signal to a particular analysis module based on a type of signal that the signal is;   third program instructions executable by a processor to cause the processer to use each analysis module to analyze one signal with a plurality of algorithms;   fourth program instructions executable by a processor to cause the processer to use each analysis module to fuse the outputs of the plurality of algorithms to provide an output;   fifth program instructions executable by a processor to cause the processer to select the highest quality output from the plurality of analysis modules; and   sixth program instructions executable by a processor to cause the processer to use the selected output to determine at least one of heart rate complexity and heart rate variability.